EvoHarnessBench: Can Your Agents Keep Pace with an Evolving Harness?New
A benchmark for agents whose tools, skills, and specialist agents keep changing — showing that harness expansion alone can make agents forget tasks they used to solve.
2026
Senior Research Scientist · Salesforce AI Research
I'm Jiayu (pronounced “Jee-ah-yü Wahng”), a Senior Research Scientist at Salesforce AI Research. I'm passionate about building efficient and intelligent agentic systems. Before Salesforce, I did my PhD in Computer Sciences at UW-Madison, advised by Prof. Aws Albarghouthi and Prof. Fred Sala (Sprocket Lab).
My recent work focuses on:
I'm always happy to discuss research, answer questions, or just chat! The best way to reach me is to open a GitHub issue, send an email — jiayuwang@salesforce.com for work-related topics or milawang0223@gmail.com for personal ones — or message me on LinkedIn.
A benchmark for agents whose tools, skills, and specialist agents keep changing — showing that harness expansion alone can make agents forget tasks they used to solve.
Learns transferable skills and per-model competence profiles to route compound AI systems — 700× cheaper to learn than Router-R1.
A live benchmark for deep research, with an evaluation suite that closely aligns with human judgment across six dimensions.
SPARKLE: what exactly does RL improve? Dissects reasoning into plan-following, problem decomposition, and knowledge use — and shows how to turn hard problems into training signal.
Accurately and efficiently predicts the performance and cost of training- and test-time adaptation strategies — smarter choices, fewer experiments.
No papers match the selected tags.
* equal contribution
Outside of research, I love playing tennis 🎾 and try to get on the court as often as I can — usually 4–5 times a week.